Definition that holds up in operations
Multi-agent AI means more than one autonomous worker participates in the same business outcome, sharing context and respecting boundaries. The hard part is coordination, not spawning extra chatbots.
Signals you need coordination
Consider multi-agent operations when: (1) work routinely crosses departments, (2) approvals must be centralized, (3) audit trails are non-negotiable, (4) integrations feed multiple functions, or (5) customers expect fast, consistent follow-through across touchpoints.
When a single agent or Tradecraft is enough
A narrow job (a weekly newsletter draft, single-system reporting) may need one specialist plus a Tradecraft, not full workforce deployment. Scale up when volume, risk, or cross-team dependencies grow.
Blitzify framing
Zed coordinates specialists; Tradecrafts supply specialized methods; Missions align objectives. Platform overview →
Frequently asked questions
Does multi-agent always mean higher C10 usage? More coordinated work can consume more capacity, but governance prevents runaway execution. Capacity planning belongs in Command Center, not guesswork.
Tradecrafts overview → | Intelligence library → | Missions →